research in natural question answering systems was first reported about 44 years ago by mccarthy and l.green (1959).

the goal of these systems was the implementation of conversational computer programs that understand and respond to queries posed in natural language. within the area of information retrieval, question answering has been associated with fact retrieval, as opposed to document retrieval. that is, it is concerned with the direct answering of questions, instead of referencing of sets of documents that might contain the answers to given questions.

simmons surveyed the existing question answering systems in 1965. Those systmes can be classified into three general apporaches to the problem: logical grammatical and computational.

Of the logical approaches SAD SAM (Lindsay) and BASEBAL (Green, 1963) allowed only very specific predicates to be used: family relations and baseball statistics, respectively. Whereas, SIR (Raphael) and SQA (Black) allowed predicates such as subset, superset, on top of, below, left of, right of, etc. in contrast Darlingotn's logic programs (1962, 1964) allowed deduction within propositional calculus. Both the DEACON (thompson) and PROSYSNTHEX (Simmons, 1964) systems used a Phrase structure Grammer approach. That is, questions were matched against data base sentences on a syntactical basis to retrieve the answer ... (to be continued. source: cognitive memory, interim report 12.8, 1969-70, coordinated science lab, university of illinios, project leader heinz von foerster)

sad sam: syntactic appraiser and diagrammer semantic analyzing machine.

The program accepts English sentences about kinship relations, builds a data base and answers questions about the facts it has stored.

SAD-SAM INPUT: John is Mary's son. SAD-SAM OUTPUT: Mary's brother is John's uncle; Mary's mother is John's grandmother, etc.

literature